In this Rmarkdown document we provide the following workflow:
Objective 0 - To examine the volume and temporal trends of existing meta-analyses on the effects of organochlorine pesticides
Objective 1: To evaluate the methodological patterns and quality of existing meta-analyses studying the effects of organochlorine pesticides.
Objective 2 . To explore the various characteristics of the organochlorine pesticides literature such as the pesticides used, the impacts elicited in response and the subjects that were investigated.
Objective 4 - To investigate the research outputs across different countries, continents and disciplines, and investigate the degree of cross-country collaboration.
Load packages and data
Load packages
rm(list = ls())
pacman::p_load(tidyverse,
hrbrthemes,
patchwork,
here,
stringr,
knitr,
formatR,
forcats,
ggplot2,
bibliometrix,
igraph,
stringi,
stringdist,
circlize,
ggalluvial,
ggraph)
knitr::opts_chunk$set(echo = TRUE, message = FALSE, warning = FALSE)
Load data
Manually extracted pilot data is stored in five separate .csv files representing different aspects of the data (extracted via structured predefined Google Forms - one per table). Bibliographic data records are exported from Scopus (including cited references field) in .bib format and locally saved as bib_sco.bib.
# Load CSV datasets
sd <- read_csv(here("data", "ocp_srm_study_details.csv"))
ocp <- read_csv(here("data", "ocp_srm_ocp_details.csv"))
sub <- read_csv(here("data", "ocp_srm_subject_details.csv"))
im <- read_csv(here("data", "ocp_srm_impact_details.csv"))
sp <- read_csv(here("data", "ocp_srm_species_details.csv"))
# Load BibTeX dataset
bib_sco <- convert2df(here("data", "bib_sco.bib"), dbsource = "scopus", format = "bibtex")
Objective 0 - To examine the volume and temporal trends of existing meta-analyses on the effects of organochlorine pesticides
Figure 1
- Bar chart showing the annual number of published meta-analysis synthesising research on the impacts of organochlorine pesticides. Divided for different subjects of exposure.
- Area graph showing the cumulative time trends of meta-analysis synthesising research on the impacts of organochlorine pesticides. Divided for different subjects of exposure.
# Join the study and subject datasets
sd_sub <- left_join(sd, sub, by = "study_id")
# Count the number of publications per year and calculate cumulative sum
sd1 <- sd %>%
count(publication_year) %>%
mutate(n_cumulative = cumsum(n))
# Create a theme for the count plots
theme_count <- function() {
theme_minimal() +
theme(
panel.grid.major.y = element_line(color = "gray", linetype = "dashed"),
panel.grid.minor.y = element_blank(),
axis.line.x = element_line(size = 1.2),
axis.line.y = element_line(size = 1.2),
axis.title.x = element_text(size = 20, face = "bold"),
axis.title.y = element_text(size = 20, face = "bold"),
axis.text.x = element_text(angle = 45, hjust = 1, size = 15, color = "#666666"),
axis.text.y = element_text(size = 20, color = "#666666"),
legend.position = "none"
)
}
# Create the annual count plot
fig1a <- sd_sub %>%
mutate(subjects = strsplit(subject, ",\\s+")) %>%
unnest(subjects) %>%
count(publication_year, subjects) %>%
group_by(subjects = reorder(subjects, n)) %>%
ggplot(aes(x = publication_year, y = n, fill = subjects)) +
geom_bar(stat = "identity", position = "stack", alpha = 0.7) +
geom_text(aes(label = n), position = position_stack(vjust = 0.6),
fontface = "bold", color = "white", size = 6, hjust = 0.4) +
scale_x_continuous(breaks = seq(min(sd1$publication_year), max(sd1$publication_year), by = 1),
expand = expansion(0,0)) +
scale_y_continuous("Annual Article Count", limits = c(0,15)) +
labs(y = "Article Count", tag = "A", fill = "Subjects") +
theme_count() +
theme(axis.title.x = element_blank()) +
scale_fill_brewer(palette = "Dark2") +
guides(fill = guide_legend(override.aes = list(size=10)))
# Create the cumulative count plot
fig1b <- sd_sub %>%
mutate(subjects = strsplit(subject, ",\\s+")) %>%
unnest(subjects) %>%
count(publication_year, subjects) %>%
group_by(subjects = reorder(subjects, n)) %>%
mutate(n_cumulative = cumsum(n)) %>%
ggplot(aes(x = publication_year, y = n_cumulative, fill = subjects)) +
geom_area(size = 1.2, alpha = 0.7) +
scale_y_continuous("Cumulative Article Count", limits = c(0,115)) +
scale_x_continuous(breaks = seq(min(sd_sub$publication_year), max(sd_sub$publication_year), by = 1),
expand = expansion(0,0)) +
labs(x = "Year", y = "Cumulative Article Count", tag = "B", fill = "Subjects") +
theme_count() +
scale_fill_brewer(palette = "Dark2") +
guides(fill = guide_legend(override.aes = list(size=10)))
# Combine the two plots
fig1 <- fig1a / fig1b
fig1
# ggsave(here("figures", "fig1.pdf"), width = 16, height = 12, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "fig1.jpg"), width = 16, height = 12, units = "cm", scale = 2, dpi = 800)
Figure s4
- Bar chart showing the annual number of published meta-analyses
synthesising research on the impacts of organochlorine pesticides.
- Line graph showing the cumulative time trends of meta-analyses synthesising research on the impacts of organochlorine pesticides.
# Create the annual Count Plot
figs4a <- sd1 %>%
ggplot(aes(x = publication_year, y = n)) +
geom_col(fill = "#1b9e77", alpha = 0.7) +
geom_text(aes(label = n), position = position_stack(vjust = 0.6),
fontface = "bold", color = "white", size = 6, hjust = 0.4) +
scale_x_continuous(breaks = seq(min(sd1$publication_year), max(sd1$publication_year), by = 1),
expand = expansion(0,1)) +
scale_y_continuous("Annual Article Count", limits = c(0,15)) +
labs(x = "Year", tag = "A") +
theme_count() +
theme(axis.title.x = element_blank())
# Create the cumulative Count Plot
figs4b <- sd1 %>%
ggplot(aes(x = publication_year, y = n_cumulative)) +
geom_line(color = "#7570b3", size = 1, linetype = "solid") +
geom_point(shape = 21, size = 6, fill = "#7570b3", stroke = 0) +
geom_text(aes(label = n_cumulative), hjust = 0.2, vjust = -1, size = 6, color = "black") +
scale_x_continuous(breaks = seq(min(sd1$publication_year), max(sd1$publication_year), by = 1),
expand = expansion(0,1)) +
scale_y_continuous("Cumulative Article Count", limits = c(0,120)) +
labs(x = "Year", tag = "B") +
theme_count()
# Combine the two plots
figs4 <- figs4a / figs4b
figs4
# ggsave(here("figures", "figs4.pdf"), width = 16, height = 12, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs4.jpg"), width = 16, height = 12, units = "cm", scale = 2, dpi = 800)
Objective 1
To evaluate the methodological patterns and quality of existing meta-analyses studying the effects of organochlorine pesticides.
Figure s5
Bar plot showing the percentage and total count of scientific literature databases used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved multiple scientific literature databases. The “Other databases” category includes all databases with a count of 3 or less.
# Calculate the total count for each category
database_count <- sd %>%
separate_rows(database_search, sep = ",\\s+") %>%
count(database_search) %>%
filter(database_search != "not reported") %>%
arrange(desc(n)) %>%
mutate(database_search = ifelse(n<= 3, "Other databases", as.character(database_search))) %>%
group_by(database_search) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
database_pct <- database_count %>%
mutate(proportion = n / sum(database_count$n),
percentage = proportion * 100)
# Create a standard theme for the supplement plots
theme_suppl <- function() {
theme_minimal() +
theme(
axis.text.y = element_text(size = 15),
axis.text.x = element_text(size = 15),
axis.line.x = element_line(color = "gray", size = 0.5),
axis.line.y = element_blank(),
axis.ticks.x = element_blank(),
axis.ticks.y = element_blank(),
axis.title.x = element_text(size = 20),
axis.title.y = element_text(size = 20),
panel.grid.major.y = element_blank(),
panel.grid.minor.y = element_blank(),
legend.position = "none"
)
}
# Create the count plot
figs5 <- database_count %>%
ggplot(aes(x = n, y = reorder(database_search, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(database_search, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = database_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(database_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs5
# ggsave(here("figures", "figs5.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs5.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s6
Alluvial plot showing the relationship between the Journal Citation Report Category and the scientific literature database used. Filtered for scientific literature database counts greater than or equal to 3.
# Rename Environmental Science
sd <- sd %>%
mutate(Journal_Category_Allocated_Broad = str_replace(Journal_Category_Allocated_Broad, "Environmental Science", "Environmental\nScience"))
# Data Transformation
database_alluvial <- sd %>%
separate_rows(database_search, sep = ",\\s+") %>%
group_by(Journal_Category_Allocated_Broad, database_search) %>%
count(database_search, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(database_search) %>%
filter(sum(freq) >= 3) %>%
filter(database_search != "NA")
# Create the Alluvial plot
figs6 <- database_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = database_search)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "Database Search"), expand = c(.05, .05)) +
xlab("Variables") +
ylab("Frequency") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/4.5, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 6) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs6
# ggsave(here("figures", "figs6.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs6.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s7
Bar plot showing the percentage and total count of effect size calculation types used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved multiple effect size calculations. The “Other effect sizes” category includes all effect sizes with a count of 2 or less.
# Calculate the total count for each category
effectsize_count <- sd %>%
separate_rows(effect_size, sep = ",\\s+") %>%
count(effect_size) %>%
filter(effect_size != "NA") %>%
mutate(effect_size = ifelse(n<= 2, "other effect sizes", as.character(effect_size))) %>%
group_by(effect_size) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
effectsize_pct <- effectsize_count %>%
mutate(proportion = n / sum(effectsize_count$n),
percentage = proportion * 100)
# Create the count plot
figs7 <- effectsize_count %>%
ggplot(aes(x = n, y = reorder(effect_size, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(effect_size, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = effectsize_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(effectsize_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs7
# ggsave(here("figures", "figs7.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs7.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s8
Alluvial plot showing the relationship between the Journal Citation Report Category and the effect size used. Filtered for scientific literature database counts greater than or equal to 3.
# Data Transformation
effectsize_alluvial <- sd %>%
separate_rows(effect_size, sep = ",\\s+") %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
filter(!grepl("no category found", Journal_Category_Allocated_Broad, ignore.case = TRUE)) %>%
filter(!is.na(effect_size)) %>%
group_by(Journal_Category_Allocated_Broad, effect_size) %>%
count(effect_size, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(effect_size)
# Create the Alluvial plot
figs8 <- effectsize_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = effect_size)) +
scale_x_discrete(limits = c("Journal Category", "Effect Size"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/2.5, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 5) +
theme_minimal() +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs8
# ggsave(here("figures", "figs8.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs8.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s9
Bar plot showing the percentage and total count of software for analysis used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved multiple software.
# Calculate the total count for each category
software_count <- sd %>%
separate_rows(software_analysis, sep = ",\\s+") %>%
mutate(software_analysis = ifelse(grepl("comprehensive meta-analysis", software_analysis), "CMAS", software_analysis)) %>%
mutate(software_analysis = ifelse(grepl("not reported", software_analysis), "no software reported", software_analysis)) %>%
count(software_analysis) %>%
filter(software_analysis != "NA") %>%
group_by(software_analysis) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
software_pct <- software_count %>%
mutate(proportion = n / sum(software_count$n),
percentage = proportion * 100)
# Create the count plot
figs9 <- software_count %>%
ggplot(aes(x = n, y = reorder(software_analysis, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(software_analysis, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = software_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(software_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs9
# ggsave(here("figures", "figs9.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs9.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s10
Alluvial plot showing the relationship between the Journal Citation Report Category and the software used for analysis.
# Data Transformation
software_analysis_alluvial <- sd %>%
separate_rows(software_analysis, sep = ",\\s+") %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
mutate(software_analysis = ifelse(grepl("comprehensive meta-analysis", software_analysis), "CMAS", software_analysis)) %>%
mutate(software_analysis = ifelse(grepl("not reported", software_analysis), "no software reported", software_analysis)) %>%
filter(!is.na(software_analysis)) %>%
group_by(Journal_Category_Allocated_Broad, software_analysis) %>%
count(software_analysis, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(software_analysis)
# Create the Alluvial plot for Software Analysis
figs10 <- software_analysis_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = software_analysis)) +
scale_x_discrete(limits = c("Journal Category", "Software Analysis"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/4, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 6) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs10
# ggsave(here("figures", "figs10.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs10.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s11
Bar plot showing the percentage and total count of heterogeneity assessment methods used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved multiple heterogeneity assessments.
# Calculate the total count for each category
heterogeneity_count <- sd %>%
separate_rows(heterogeneity_assessment_method, sep = ",\\s+") %>%
count(heterogeneity_assessment_method) %>%
mutate(heterogeneity_assessment_method = ifelse(grepl("not reported", heterogeneity_assessment_method), "no heterogeneity\nmeasure reported", heterogeneity_assessment_method)) %>%
filter(heterogeneity_assessment_method != "NA") %>%
group_by(heterogeneity_assessment_method) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
heterogeneity_pct <- heterogeneity_count %>%
mutate(proportion = n / sum(heterogeneity_count$n),
percentage = proportion * 100)
# Create the count plot
figs11 <- heterogeneity_count %>%
ggplot(aes(x = n, y = reorder(heterogeneity_assessment_method, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(heterogeneity_assessment_method, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = heterogeneity_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(heterogeneity_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs11
# ggsave(here("figures", "figs11.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs11.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s12
Alluvial plot showing the relationship between the Journal Citation Report Category and heterogeneity assessment method.
# Data Transformation
heterogeneity_alluvial <- sd %>%
separate_rows(heterogeneity_assessment_method, sep = ",\\s+") %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
mutate(heterogeneity_assessment_method = ifelse(grepl("not reported", heterogeneity_assessment_method), "no heterogeneity\nmeasure reported", heterogeneity_assessment_method)) %>%
filter(!is.na(heterogeneity_assessment_method)) %>%
group_by(Journal_Category_Allocated_Broad, heterogeneity_assessment_method) %>%
count(heterogeneity_assessment_method, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(heterogeneity_assessment_method)
# Create the Alluvial plot
figs12 <- heterogeneity_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = heterogeneity_assessment_method)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "Heterogeneity"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/4.5, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 6) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs12
# ggsave(here("figures", "figs12.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs12.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s13
Bar plot showing the percentage and total count of sensitivity analyses conducted in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved multiple sensitivity analyses.
# Calculate the total count for each category
sensitivity_count <- sd %>%
separate_rows(sensitivity_analysis_method, sep = ",\\s+") %>%
count(sensitivity_analysis_method) %>%
filter(sensitivity_analysis_method != "NA") %>%
group_by(sensitivity_analysis_method) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
sensitivity_pct <- sensitivity_count %>%
mutate(proportion = n / sum(sensitivity_count$n),
percentage = proportion * 100)
# Create the count plot
figs13 <- sensitivity_count %>%
ggplot(aes(x = n, y = reorder(sensitivity_analysis_method, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(sensitivity_analysis_method, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = sensitivity_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(sensitivity_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs13
# ggsave(here("figures", "figs13.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs13.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s14
Alluvial plot showing the relationship between the Journal Citation Report Category and sensitivity analysis.
# Data Transformation
sensitivity_analysis_alluvial <- sd %>%
separate_rows(sensitivity_analysis_method, sep = ",\\s+") %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
filter(!is.na(sensitivity_analysis_method)) %>%
group_by(Journal_Category_Allocated_Broad, sensitivity_analysis_method) %>%
count(sensitivity_analysis_method, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(sensitivity_analysis_method)
# Create the Alluvial plot
figs14 <- sensitivity_analysis_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = sensitivity_analysis_method)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "Sensitivity Analysis Method"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/4, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 5) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs14
# ggsave(here("figures", "figs14.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs14.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s15
Bar plot showing the percentage and total count of bias types assessed in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved the assessment of multiple bias types.
# Calculate the total count for each category
bias_type_count <- sd %>%
separate_rows(bias_assessment_type, sep = ",\\s+") %>%
count(bias_assessment_type) %>%
filter(bias_assessment_type != "NA") %>%
group_by(bias_assessment_type) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
bias_type_pct <- bias_type_count %>%
mutate(proportion = n / sum(bias_type_count$n),
percentage = proportion * 100)
# Create the count plot
figs15 <- bias_type_count %>%
ggplot(aes(x = n, y = reorder(bias_assessment_type, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(bias_assessment_type, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = bias_type_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(bias_type_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs15
# ggsave(here("figures", "figs15.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs15.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s16
Alluvial plot showing the relationship between the Journal Citation Report Category and bias assessment method
# Data Transformation
bias_assessment_alluvial <- sd %>%
separate_rows(bias_assessment_type, sep = ",\\s+") %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
filter(!is.na(bias_assessment_type)) %>%
group_by(Journal_Category_Allocated_Broad, bias_assessment_type) %>%
count(bias_assessment_type, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(bias_assessment_type)
# Create the Alluvial plot
figs16 <- bias_assessment_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = bias_assessment_type)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "Bias Assessment Type"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/4, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 6) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs16
# ggsave(here("figures", "figs16.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs16.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s17
Bar plot showing the percentage and total count of bias methodologies used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved the use of multiple bias assessment methodologies.
# Calculate the total count for each category
bias_method_count <- sd %>%
separate_rows(bias_assessment_method, sep = ",\\s+") %>%
count(bias_assessment_method) %>%
filter(bias_assessment_method != "NA") %>%
group_by(bias_assessment_method) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
bias_method_pct <- bias_method_count %>%
mutate(proportion = n / sum(bias_method_count$n),
percentage = proportion * 100)
# Create the count plot
figs17 <- bias_method_count %>%
ggplot(aes(x = n, y = reorder(bias_assessment_method, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(bias_assessment_method, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = bias_method_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(bias_method_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs17
# ggsave(here("figures", "figs17.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs17.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s18
Alluvial plot showing the relationship between the Journal Citation Report Category and bias assessment method.
# Data Transformation
bias_assessment_alluvial <- sd %>%
separate_rows(bias_assessment_method, sep = ",\\s+") %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
filter(!is.na(bias_assessment_method)) %>%
group_by(Journal_Category_Allocated_Broad, bias_assessment_method) %>%
count(bias_assessment_method, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(bias_assessment_method)
# Create the Alluvial plot
figs18 <- bias_assessment_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = bias_assessment_method)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "Bias Assessment Method"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/3.5, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 6) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs18
# ggsave(here("figures", "figs18.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs18.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s19
Bar plot showing the percentage and total count of bias visualizations used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved the use of multiple bias visualizations.
# Calculate the total count for each category
bias_visualization_count <- sd %>%
separate_rows(bias_assessment_visualization, sep = ",\\s+") %>%
count(bias_assessment_visualization) %>%
filter(bias_assessment_visualization != "NA") %>%
group_by(bias_assessment_visualization) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
bias_visualization_pct <- bias_visualization_count %>%
mutate(proportion = n / sum(bias_visualization_count$n),
percentage = proportion * 100)
# Create the count plot
figs19 <- bias_visualization_count %>%
ggplot(aes(x = n, y = reorder(bias_assessment_visualization, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(bias_assessment_visualization, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = bias_visualization_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(bias_visualization_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs19
# ggsave(here("figures", "figs19.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs19.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s20
Alluvial plot showing the relationship between the Journal Citation Report Category and bias visualization method.
# Data Transformation
bias_vizualisation_alluvial <- sd %>%
separate_rows(bias_assessment_visualization, sep = ",\\s+") %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
filter(!is.na(bias_assessment_visualization)) %>%
group_by(Journal_Category_Allocated_Broad, bias_assessment_visualization) %>%
count(bias_assessment_visualization, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(bias_assessment_visualization)
# Create the Alluvial plot
figs20 <- bias_vizualisation_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = bias_assessment_visualization)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "Bias Assessment Method"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/4, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 6) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs20
# ggsave(here("figures", "figs20.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs20.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s21
Bar plot showing the percentage and total count of risk of bias tests used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved the use of multiple risk of bias tests.
# Calculate the total count for each category
rob_method_count <- sd %>%
separate_rows(rob_assessment_method, sep = ",\\s+") %>%
count(rob_assessment_method) %>%
filter(rob_assessment_method != "NA") %>%
group_by(rob_assessment_method) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
rob_method_pct <- rob_method_count %>%
mutate(proportion = n / sum(rob_method_count$n),
percentage = proportion * 100)
# Create the count plot
figs21 <- rob_method_count %>%
ggplot(aes(x = n, y = reorder(rob_assessment_method, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(rob_assessment_method, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = rob_method_pct, aes(label = paste0("(", round(percentage, 0), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(rob_method_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs21
# ggsave(here("figures", "figs21.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs21.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s22
Alluvial plot showing the relationship between the Journal Citation Report Category and risk of bias methodology.
# Data Transformation
robmethod_alluvial <- sd %>%
separate_rows(rob_assessment_method, sep = ",\\s+") %>%
filter(!is.na(rob_assessment_method)) %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
group_by(Journal_Category_Allocated_Broad, rob_assessment_method) %>%
count(rob_assessment_method, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(rob_assessment_method)
# Create the Alluvial plot
figs22 <- robmethod_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = rob_assessment_method)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "ROB Assessment Method"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/3, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 5) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs22
# ggsave(here("figures", "figs22.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs22.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s23
Bar plot showing the percentage and total count of visualization methods used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved the use of multiple visualization methods.
# Calculate the total count for each category
visualization_count <- sd %>%
separate_rows(visualization_method, sep = ",\\s+") %>%
count(visualization_method) %>%
filter(visualization_method != "NA") %>%
group_by(visualization_method) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
visualization_pct <- visualization_count %>%
mutate(proportion = n / sum(visualization_count$n),
percentage = proportion * 100)
# Create the count plot
figs23 <- visualization_count %>%
ggplot(aes(x = n, y = reorder(visualization_method, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(visualization_method, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = visualization_pct, aes(label = paste0("(", round(percentage, 0), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(visualization_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs23
# ggsave(here("figures", "figs23.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs23.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s24
Alluvial plot showing the relationship between the Journal Citation Report Category and visualization method.
# Data Transformation
vizualization_alluvial <- sd %>%
separate_rows(visualization_method, sep = ",\\s+") %>%
filter(!is.na(visualization_method)) %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
group_by(Journal_Category_Allocated_Broad, visualization_method) %>%
count(visualization_method, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(visualization_method)
# Create the Alluvial plot
figs24 <- vizualization_alluvial %>%
ggplot(aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = visualization_method)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "ROB Assessment Method"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/4, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 6) +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
theme_suppl()
figs24
# ggsave(here("figures", "figs24.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs24.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s25
Bar plot showing the percentage and total count of reporting guidelines used in meta-analyses investigating the impacts of organochlorine pesticides. Note that some meta-analyses may contribute to multiple sections if the study involved the use of multiple reporting guidelines.
# Calculate the total count for each category
reporting_guide_count <- sd %>%
separate_rows(reporting_standards_type, sep = ",\\s+") %>%
count(reporting_standards_type) %>%
filter(reporting_standards_type != "NA") %>%
group_by(reporting_standards_type) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
reporting_guide_pct <- reporting_guide_count %>%
mutate(proportion = n / sum(reporting_guide_count$n),
percentage = proportion * 100)
# Create the count plot
figs25 <- reporting_guide_count %>%
ggplot(aes(x = n, y = reorder(reporting_standards_type, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(reporting_standards_type, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = reporting_guide_pct, aes(label = paste0("(", round(percentage, 0), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(reporting_guide_count$n)*1.1)) +
labs(y = NULL) +
theme_suppl()
figs25
# ggsave(here("figures", "figs25.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs25.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s26
An alluvial plot showing the relationship between the Journal Citation Report Category and reporting guideline used
# Data Transformation
reporting_standards_alluvial <- sd %>%
separate_rows(reporting_standards_type, sep = ",\\s+") %>%
separate_rows(Journal_Category_Allocated_Broad, sep = "/\\s+") %>%
filter(!grepl("no category found", Journal_Category_Allocated_Broad, ignore.case = TRUE)) %>%
filter(!is.na(reporting_standards_type)) %>%
group_by(Journal_Category_Allocated_Broad, reporting_standards_type) %>%
count(reporting_standards_type, Journal_Category_Allocated_Broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(reporting_standards_type)
# Create the Alluvial plot
figs26 <- reporting_standards_alluvial %>%
ggplot( aes(y = freq ,axis1 = Journal_Category_Allocated_Broad, axis2 = reporting_standards_type)) +
scale_x_discrete(limits = c("Journal Citation Report Category", "Reporting Standards Type"), expand = c(.05, .05)) +
xlab("Variables") +
geom_alluvium(aes(fill = Journal_Category_Allocated_Broad)) +
geom_stratum(width = 1/3.5, fill = "white", color = "black") +
labs(x= "Variables", y = "Frequency", fill = "Journal Category Allocated") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 6) +
theme_suppl()
figs26
# ggsave(here("figures", "figs26.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs26.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure 2
A circular treemap showing the counts of each methodological item in exisitng meta-analysis investigating the impacts of organochlorine pesticides
# Grouping "Medline" and "Pubmed" under "PubMed" and summing the counts
database_count <- database_count %>%
mutate(database_search = if_else(database_search %in% c("Medline", "Pubmed"), "PubMed", database_search)) %>%
group_by(database_search) %>%
summarise(n = sum(n))
# Splitting, grouping, and summing different categories of effect sizes
effectsize_count <- sd %>%
separate_rows(effect_size, sep = ",\\s+") %>%
count(effect_size) %>%
filter(effect_size != "NA") %>%
group_by(effect_size) %>%
summarise(n = sum(n))
effectsize_count <- effectsize_count %>%
mutate(effect_size = if_else(effect_size %in% c("Beta regression coefficient", "correlation coefficient"), "Correlation", effect_size)) %>%
mutate(effect_size = if_else(effect_size %in% c("odds ratio", "response ratio", "SMD (standardized mean difference)", "lnPR (log partitioning ratio)", "lnOR (log odds ratio)", "ratio of means", "InRR (log response ratio)"), "Mean diff", effect_size)) %>%
mutate(effect_size = if_else(effect_size %in% c("risk ratio"), "2x2", effect_size)) %>%
mutate(effect_size = if_else(effect_size %in% c("raw weight", "transfer rate", "geometric mean", "maternal transfer ratio", "standardized mortality rate", "Transfer rate", "z-score", "lnCVR (log coefficient variation ratio)"), "Other ES", effect_size)) %>%
group_by(effect_size) %>%
summarise(n = sum(n))
# Categorizing software as either "Code-based software" or "GUI" and summing the counts
software_count <- software_count %>%
mutate(software_analysis = if_else(software_analysis %in% c("Stata", "R"), "Code-based", software_analysis)) %>%
mutate(software_analysis = if_else(software_analysis %in% c("CMAS", "Excel", "RevMan", "SAS", "XLSTAT"), "GUI", software_analysis)) %>%
group_by(software_analysis) %>%
summarise(n = sum(n))
# Grouping heterogeneity assessment methods and summing the counts
heterogeneity_count <- heterogeneity_count %>%
mutate(heterogeneity_assessment_method = if_else(heterogeneity_assessment_method %in% c("Tau square"), "I squared", heterogeneity_assessment_method)) %>%
mutate(heterogeneity_assessment_method = if_else(heterogeneity_assessment_method %in% c("Chi square"), "Q statistic", heterogeneity_assessment_method)) %>%
group_by(heterogeneity_assessment_method) %>%
summarise(n = sum(n))
# Grouping bias assessment methods with count <= 3 under "Other BA" and summing the counts
bias_method_count <- bias_method_count %>%
mutate(bias_assessment_method = if_else(bias_assessment_method %in% c("Egger's regression test", "Fail safe", "Begg's test", "Kendall's tau statistic"),"Bias statistical test", bias_assessment_method)) %>%
group_by(bias_assessment_method) %>%
summarise(n = sum(n))
# Grouping sensitivity analysis methods and summing the counts
sensitivity_count <- sensitivity_count %>%
mutate(sensitivity_analysis_method = ifelse(n<= 8, "Other sensitivity analysis", as.character(sensitivity_analysis_method))) %>%
group_by(sensitivity_analysis_method) %>%
summarise(n = sum(n))
# Grouping risk of bias assessment methods with count <= 3 under "Other ROB" and summing the counts
rob_method_count <- rob_method_count %>%
mutate(rob_assessment_method = ifelse(n<= 3, "Other ROB", as.character(rob_assessment_method))) %>%
group_by(rob_assessment_method) %>%
summarise(n = sum(n))
# Grouping visualization methods with count <= 3 under "Other Viz" and summing the counts
visualization_count <- visualization_count %>%
mutate(visualization_method = ifelse(n<= 3, "Other Viz", as.character(visualization_method))) %>%
group_by(visualization_method) %>%
summarise(n = sum(n))
# Grouping reporting standards types with count <= 2 under "Other guideline" and summing the counts
reporting_guide_count <- reporting_guide_count %>%
mutate(reporting_standards_type = ifelse(n<= 2, "Other guideline", as.character(reporting_standards_type))) %>%
group_by(reporting_standards_type) %>%
summarise(n = sum(n))
# Grouping bias assessment visualization types and summing the counts
bias_visualization_count <- bias_visualization_count %>%
mutate(bias_assessment_visualization = if_else(bias_assessment_visualization %in% c("Doi plot", "Trim and fill", "Funnel plot", "Galbraith plot"), "Bias visualization", bias_assessment_visualization)) %>%
group_by(bias_assessment_visualization) %>%
summarise(n = sum(n))
# Combine the data frames and unite the methodology types
df <- bind_rows(
database_count %>% mutate(methodology_type = 'Database Search'),
effectsize_count %>% mutate(methodology_type = 'Effect Size'),
software_count %>% mutate(methodology_type = "Software"),
heterogeneity_count %>% mutate(methodology_type = "Heterogeneity"),
sensitivity_count %>% mutate(methodology_type = "Sensitivity_Analysis"),
bias_method_count %>% mutate(methodology_type = "Bias Assessment"),
rob_method_count %>% mutate(methodology_type = "Risk of Bias"),
# visualization_count %>% mutate(methodology_type = "Visualization"),
reporting_guide_count %>% mutate(methodology_type = "Reporting Guide"),
bias_visualization_count %>% mutate(methodology_type = "Bias Assessment")
) %>%
unite(methodology_type_specific,
database_search,
effect_size,
software_analysis,
heterogeneity_assessment_method,
sensitivity_analysis_method,
bias_assessment_method,
rob_assessment_method,
# visualization_method,
reporting_standards_type,
bias_assessment_visualization,
remove = TRUE, na.rm = TRUE)
# Preparing the edges dataframe for creating the graph
edges <- df %>%
rename(from = methodology_type, to = methodology_type_specific, size = n) %>%
select(c(from,to,size)) %>%
as.data.frame()
# Preparing the vertices dataframe for creating the graph
vertices <- df %>%
rename(name = methodology_type_specific, size = n) %>%
select(c(name,size)) %>%
as.data.frame()
# Appending unique 'from' values to vertices and their corresponding summed sizes
vertices[(nrow(edges)+1):(nrow(edges)+length(unique(edges$from))),1] <- unique(edges$from)
N <- aggregate(edges$size, list(edges$from), FUN=sum)
vertices[(nrow(edges)+1):(nrow(edges)+length(unique(edges$from))),2] <- N$x
# Creating a graph object from the edges and vertices dataframes
mygraph <- graph_from_data_frame(edges, vertices = vertices)
# Plotting the graph using a 'circlepack' layout, adding labels and adjusting aesthetics
fig2 <- ggraph(mygraph, layout = 'circlepack', weight = size) +
geom_node_circle(aes(fill = as.factor(depth)), color = NA) +
scale_fill_manual(values = c("#1b9e77", "#88D1AD")) +
geom_node_text(aes(label = name, filter = leaf, size = 10), vjust = -0.3, fontface = "bold") +
geom_node_text(aes(label = paste0("(", size, ")"), filter = leaf, size = 10), vjust = 1, fontface = "bold") +
theme_void() +
theme(legend.position = "none")
fig2
ggsave(here("figures", "fig2.pdf"), width = 18, height = 15, units = "cm", scale = 2, dpi = 800)
ggsave(here("figures", "fig2.jpg"), width = 18, height = 15, units = "cm", scale = 2, dpi = 800)
Figure 3
The average reporting quality and rigour of meta-analysis according to CEESAT 2.1 (Woodcock et al., 2014). Gold is regard as the highest score, green is second highest score, amber is second-lowest score, and red is the lowest score. All CEESAT 2.1 items along with our interpretation are provided in the Supplementary File 2.
# Start the data manipulation
percent_ceesat_score <- sd %>%
filter(!is.na(author_year)) %>%
select(studies = author_year, starts_with("CEE")) %>%
na.omit() %>%
pivot_longer(cols = -studies, names_to = "question", values_to = "score") %>%
group_by(question, score) %>%
summarise(n = n(), .groups = 'drop') %>%
mutate(percent = (n/sum(n))*100,
across(c(question, score), as.factor),
question = fct_recode(question,
`1.1 Are the elements of the review question clear?` = "CEESAT2_1.1",
`2.1 Is there an a-priori method protocol document?` = "CEESAT2_2.1",
`3.1. Is the approach to searching clearly definedsystematic and\ntransparent?` = "CEESAT2_3.1",
`3.2. Is the search comprehensive?` = "CEESAT2_3.2",
`4.1. Are eligibility criteria clearly defined?` = "CEESAT2_4.1",
`4.2. Are eligibility criteria consistently applied to all potentially relevant\narticles and studies found during the search?` = "CEESAT2_4.2",
`4.3. Are eligibility decisions transparently reported?` = "CEESAT2_4.3",
`5.1. Does the review critically appraise each study?` = "CEESAT2_5.1",
`5.2. During critical appraisal was an effort made to minimise\nsubjectivity?` = "CEESAT2_5.2",
`6.1. Is the method of data extraction fully documented?` = "CEESAT2_6.1",
`6.2. Are the extracted data reported for each study?` = "CEESAT2_6.2",
`6.3. Were extracted data cross checked by more than one reviewer?` = "CEESAT2_6.3",
`7.1. Is the choice of synthesis approach appropriate?` = "CEESAT2_7.1",
`7.2. Is a statistical estimate of pooled effect provided together with\nmeasure of variance and heterogeneity among studies?` = "CEESAT2_7.2",
`7.3 Is variability in the study findings investigated and discussed?` = "CEESAT2_7.3",
`8.1 Have the authors considered limitations in the synthesis?` = "CEESAT2_8.1"),
question = factor(question, levels = rev(levels(question))),
score = factor(score, levels = levels(score)[c(4,1,3,2)]))
# Create CEESAT plot
fig3 <- ggplot(data = percent_ceesat_score, aes(x = question, y = percent, fill = score)) +
geom_col(width = 0.7, position = "fill", color = "black") +
geom_text(aes(label = n), position = position_fill(vjust = 0.5), size = 7, fontface = "bold") +
coord_flip() +
guides(fill = guide_legend(reverse = TRUE)) +
scale_fill_manual(values = c("#FF0000","#FFD700","#008000", "#DAA520"), name = "Score:") +
scale_y_continuous(labels = scales::percent) +
theme(panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_blank(),
axis.text.y = element_text(size = 20),
axis.text.x = element_text(size = 20),
axis.title.x = element_text(size = 25),
axis.title.y = element_text(size = 25),
legend.position = "none") +
ylab("Percentage") +
xlab("CEESAT Question")
fig3
# ggsave(here("figures", "fig3.pdf"), width = 25, height = 15, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "fig3.jpg"), width = 25, height = 15, units = "cm", scale = 2, dpi = 800)
Objective 2
To explore the various characteristics of the organochlorine pesticides literature such as the pesticides used, the impacts elicited in response and the subjects that were investigated.
Figure s27
A bar plot showing the percentage and total count of total of pesticides investigate in meta-analysis investigating the impacts of organochlorine pesticides. Note: some meta-analysis may contribute to multiple sections if the study involves multiple organochlorine pesticides. Filtered for pesticide counts greater than 6.
# Function to replace long OCP names with abbreviations
replace_ocp <- function(df) {
df %>%
mutate(ocp = case_when(
# HCH related replacements
grepl("Hexachlorocyclohexane \\(HCH\\)", ocp, ignore.case = TRUE) ~ "HCH",
grepl("alpha-Hexachlorocyclohexane \\(alpha-HCH\\)", ocp, ignore.case = TRUE) ~ "α - HCH",
grepl("beta-Hexachlorocyclohexane \\(beta-HCH\\)", ocp, ignore.case = TRUE) ~ "β - HCH",
grepl("gamma-Hexachlorocyclohexane \\(gamma-HCH\\)", ocp, ignore.case = TRUE) ~ "Lindane",
# DDT related replacements
grepl("Dichlorodiphenyltrichloroethane \\(DDT\\)", ocp, ignore.case = TRUE) ~ "DDT",
grepl("p,p-Dichlorodiphenyltrichloroethane \\(p,p-DDT\\)", ocp, ignore.case = TRUE) ~ "p,p-DDT",
grepl("o,p-Dichlorodiphenyltrichloroethane \\(o,p-DDT\\)", ocp, ignore.case = TRUE) ~ "o,p-DDT",
# DDD related replacements
grepl("Dichlorodiphenyldichloroethane \\(DDD\\)", ocp, ignore.case = TRUE) ~ "DDD",
grepl("p,p-Dichlorodiphenyldichloroethane \\(p,p-DDD\\)", ocp, ignore.case = TRUE) ~ "p,p-DDD",
grepl("o,p-Dichlorodiphenyldichloroethane \\(o,p-DDD\\)", ocp, ignore.case = TRUE) ~ "o,p-DDD",
# DDE related replacements
grepl("Dichlorodiphenyldichloroethylene \\(DDE\\)", ocp, ignore.case = TRUE) ~ "DDE",
grepl("p,p-Dichlorodiphenyldichloroethylene \\(p,p-DDE\\)", ocp, ignore.case = TRUE) ~ "p,p-DDE",
grepl("o,p-Dichlorodiphenyldichloroethylene \\(o,p-DDE\\)", ocp, ignore.case = TRUE) ~ "o,p-DDE",
TRUE ~ ocp # no change for any others
))
}
# Transform the data
ocp_count <-
ocp %>%
separate_rows(ocp, sep = ",\\s+") %>%
replace_ocp %>%
count(ocp) %>%
filter(!is.na(ocp)) %>% # filter out NA
arrange(desc(n)) %>%
mutate(ocp = ifelse(n <= 6, "other OCP", as.character(ocp))) %>%
# Other organochlorine pesticides include: toxaphene (n=6), Methoxychlor (n=6), p,p-DDD (n=5), o,p-DDD (n=5), cis-Chlordane (n=5), trans=Chlordane (n=4), Endosulfan II (n=4), Endosulfan (n=3),o,p-DDD (n=3), cis-Nonachlor (n=3) y-Chlordane (n=2), Chlorophenol (n=2), beta-BHC (n=2), alpha-Chlordane (n=2), TCDD (n=1), Nonachlore (n=1), Endrin ketone (n=1), Endrin Aldehyde (n=1), Endosulfan sulfate (n=1), Dicofol (n=1), delta-HCH (n=1), delta-BHC (n=1), DDD (n=1), cis-Heptachlor (n=1), alpha-BHC (n=1).
group_by(ocp) %>%
summarise(n =sum(n))
# Calculate the proportion and percentage of each OCP
ocp_pct <- ocp_count %>%
mutate(proportion = n/sum(ocp_count$n),
percentage = proportion*100)
# Create the count plot for OCPs
figs27 <- ocp_count %>%
ggplot(aes(x = n, y = reorder(ocp, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8, alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(ocp, n)), hjust = 0.5, size = 6, color = "black") +
geom_text(data = ocp_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = - 0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(ocp_count$n)*1.2)) +
labs(y = NULL) +
theme_minimal() +
theme(axis.text.y = element_text(size = 15),
axis.text.x = element_text(size = 15),
axis.line.x = element_line(color = "gray", size = 0.5),
axis.line.y = element_blank(),
axis.ticks.x = element_blank(),
axis.ticks.y = element_blank(),
axis.title.x = element_text(size = 20),
axis.title.y = element_text(size = 20),
panel.grid.major.y = element_blank(),
panel.grid.minor.y = element_blank(),
legend.position = "none"
)
figs27
# ggsave(here("figures", "figs27.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs27.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s28
A bar plot showing the percentage and total count of total of subjects investigate in meta-analysis investigating the impacts of organochlorine pesticides. Note: some meta-analysis may contribute to multiple sections if the study involves multiple subjects
# Calculate total count for each category
subject_count <-
sub %>%
separate_rows(subject, sep = ",\\s+") %>%
count(subject)
# Calculate proportion and percentage for each category
subject_pct <- subject_count %>%
mutate(proportion = n/sum(subject_count$n),
percentage = proportion*100)
# Create the count plot for subjects
figs28 <- subject_count %>%
ggplot(aes(x = n, y = reorder(subject, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(subject, n)), hjust = 0.5, size = 7, color = "black") +
geom_text(data = subject_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(subject_count$n)*1.1)) +
labs(y = NULL) +
theme_minimal() +
theme(axis.text.y = element_text(size = 15),
axis.text.x = element_text(size = 15),
axis.line.x = element_line(color = "gray", size = 0.5),
axis.line.y = element_blank(),
axis.ticks.x = element_blank(),
axis.ticks.y = element_blank(),
axis.title.x = element_text(size = 20),
axis.title.y = element_text(size = 20),
panel.grid.major.y = element_blank(),
panel.grid.minor.y = element_blank(),
legend.position = "none"
)
figs28
# ggsave(here("figures", "figs28.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs28.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s29
A bar plot showing the percentage and total count of total of subjects investigate in meta-analysis investigating the impacts of organochlorine pesticides. Note: some meta-analysis may contribute to multiple sections if the study involves multiple subjects. Filtered for impact counts greater than 1.
# Calculate total count for each category
impact_count <-
im %>%
separate_rows(impact, sep = ",\\s+") %>%
count(impact) %>%
filter(impact != "NA") %>%
mutate(impact = ifelse(n<= 1, "other", as.character(impact))) %>%
group_by(impact) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
impact_pct <- impact_count %>%
mutate(proportion = n/sum(impact_count$n),
percentage = proportion*100)
# Create the count plot for impacts
figs29 <- impact_count %>%
ggplot(aes(x = n, y = reorder(impact, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(impact, n)), hjust = 0.5, size = 7, color = "black") +
geom_text(data = impact_pct, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(impact_count$n)*1.1)) +
labs(y = NULL) +
theme_minimal() +
theme(axis.text.y = element_text(size = 15),
axis.text.x = element_text(size = 15),
axis.line.x = element_line(color = "gray", size = 0.5),
axis.line.y = element_blank(),
axis.ticks.x = element_blank(),
axis.ticks.y = element_blank(),
axis.title.x = element_text(size = 20),
axis.title.y = element_text(size = 20),
panel.grid.major.y = element_blank(),
panel.grid.minor.y = element_blank(),
legend.position = "none"
)
figs29
# ggsave(here("figures", "figs29.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs29.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s30
A bar plot showing the percentage and total count of total of impact categories investigated in meta-analysis investigating the impacts of organochlorine pesticides. Note: some meta-analysis may contribute to multiple sections if the study involves multiple impacts
# Create a collumn for broad impacts
im <- im %>%
separate_rows(impact, sep = ",\\s+") %>%
mutate(impact_broad = case_when(
impact %in% c("parkinsons disease", "alzheimers disease", "autism spectrum disorder", "brain tumour", "amyotrophic lateral sclerosis" ) ~ "Neurological",
impact %in% c("concentration", "contamination") ~ "Concentration",
impact %in% c("diabetes", "thyroid function", "hypertension", "endometriosis") ~ "Endocrine",
grepl("cancer", impact, ignore.case = TRUE) | impact %in% c("leukemia", "lymphoma", "multiple myeloma", "neuroblastoma") ~ "Carcinogen",
impact %in% c("respiratory health", "cardiovascular disease", "asthma", "prolonged bradycardia") ~ "Cardiovascular",
# impact %in% c("birth outcomes", "birth weight", "preterm birth") ~ "Birth",
impact %in% c("obesity", "adiposity") ~ "Obesity",
impact %in% c("sperm quality", "neuroblastoma", "hypospadias", "cryptochidism", "reproductive system") ~ "Reproduction",
TRUE ~ "Other Impact"
))
# Calculate total count for each category
impact_count_broad <-
im %>%
separate_rows(impact_broad, sep = ",\\s+") %>%
count(impact_broad) %>%
filter(impact_broad != "NA") %>%
group_by(impact_broad) %>%
summarise(n = sum(n))
# Calculate proportion and percentage for each category
impact_pct_broad <- impact_count_broad %>%
mutate(proportion = n/sum(impact_count_broad$n),
percentage = proportion*100)
# Create the count plot for impacts
figs30 <- impact_count_broad %>%
ggplot(aes(x = n, y = reorder(impact_broad, n), fill = "#1b9e77")) +
geom_bar(stat = "identity", width = 0.8 , alpha = 0.7) +
geom_text(aes(label = n, x = n / 2, y = reorder(impact_broad, n)), hjust = 0.5, size = 7, color = "black") +
geom_text(data = impact_pct_broad, aes(label = paste0("(", round(percentage, 1), "%)"), x = n),
hjust = -0.1, size = 6, color = "black", fontface = "bold") +
scale_fill_identity(guide = "none") +
scale_x_continuous(name = "Article Count", expand = c(0, 0), limits = c(0, max(impact_count_broad$n)*1.1)) +
labs(y = NULL) +
theme_minimal() +
theme(axis.text.y = element_text(size = 15),
axis.text.x = element_text(size = 15),
axis.line.x = element_line(color = "gray", size = 0.5),
axis.line.y = element_blank(),
axis.ticks.x = element_blank(),
axis.ticks.y = element_blank(),
axis.title.x = element_text(size = 20),
axis.title.y = element_text(size = 20),
panel.grid.major.y = element_blank(),
panel.grid.minor.y = element_blank(),
legend.position = "none"
)
figs30
# ggsave(here("figures", "figs30.pdf"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs30.jpg"), width = 16, height = 10, units = "cm", scale = 2, dpi = 800)
Figure s31
An alluvial plot showing the relationships between the pesticide of exposure, the subject being exposed and the impact of exposure
# Function to change OCP to be more general
replace_ocp2 <- function(df) {
df %>%
mutate(ocp = case_when(
grepl("Chlordane", ocp, ignore.case = TRUE) ~ "Chlordane",
grepl("Endosulfan", ocp, ignore.case = TRUE) ~ "Endosulfan",
grepl("Nonachlor", ocp, ignore.case = TRUE) ~ "Nonachlore",
grepl("Heptachlor", ocp, ignore.case = TRUE) ~ "Heptachlor",
grepl("TCDD", ocp, ignore.case = TRUE) ~ "TCDD",
grepl("Endrin", ocp, ignore.case = TRUE) ~ "Endrin",
grepl("Hexachlorobenzene", ocp, ignore.case = TRUE) ~ "HCH",
grepl("Lindane", ocp, ignore.case = TRUE) ~ "HCH",
grepl("Hexachlorocyclohexane \\(HCH\\)", ocp, ignore.case = TRUE) ~ "HCH",
grepl("alpha-Hexachlorocyclohexane \\(alpha-HCH\\)", ocp, ignore.case = TRUE) ~ "HCH",
grepl("beta-Hexachlorocyclohexane \\(beta-HCH\\)", ocp, ignore.case = TRUE) ~ "HCH",
grepl("gamma-Hexachlorocyclohexane \\(gamma-HCH\\)", ocp, ignore.case = TRUE) ~ "HCH",
grepl("Dichlorodiphenyltrichloroethane \\(DDT\\)", ocp, ignore.case = TRUE) ~ "DDT",
grepl("p,p-Dichlorodiphenyltrichloroethane \\(p,p-DDT\\)", ocp, ignore.case = TRUE) ~ "DDT",
grepl("o,p-Dichlorodiphenyltrichloroethane \\(o,p-DDT\\)", ocp, ignore.case = TRUE) ~ "DDT",
grepl("Dichlorodiphenyldichloroethane \\(DDD\\)", ocp, ignore.case = TRUE) ~ "DDD",
grepl("p,p-Dichlorodiphenyldichloroethane \\(p,p-DDD\\)", ocp, ignore.case = TRUE) ~ "DDD",
grepl("o,p-Dichlorodiphenyldichloroethane \\(o,p-DDD\\)", ocp, ignore.case = TRUE) ~ "DDD",
grepl("Dichlorodiphenyldichloroethylene \\(DDE\\)", ocp, ignore.case = TRUE) ~ "DDE",
grepl("p,p-Dichlorodiphenyldichloroethylene \\(p,p-DDE\\)", ocp, ignore.case = TRUE) ~ "DDE",
grepl("o,p-Dichlorodiphenyldichloroethylene \\(o,p-DDE\\)", ocp, ignore.case = TRUE) ~ "DDE",
TRUE ~ ocp
))
}
# Transform the data
alluvial <- im %>%
left_join(ocp, by = "study_id") %>%
left_join(sub, by = "study_id") %>%
separate_rows(subject, sep = ",\\s+") %>%
separate_rows(ocp, sep = ",\\s+") %>%
separate_rows(impact_broad, sep = ",\\s+") %>%
replace_ocp2() %>%
filter(!grepl("not reported", ocp, ignore.case = TRUE)) %>%
group_by(ocp, subject, impact_broad) %>%
summarise(freq = n(), .groups = 'drop') %>%
group_by(ocp) %>%
filter(sum(freq) > 10) %>%
group_by(impact_broad) %>%
filter(sum(freq) > 5) %>%
mutate(subject = factor(subject, levels = c("Environment", "Non-human animal", "Human"), ordered = TRUE))
# Make alluvial plot
figs31 <- alluvial %>%
ggplot(
aes(axis1 = ocp, axis2 = subject, axis3 = impact_broad, y = freq)) +
scale_x_discrete(limits = c("Organochlorine Pesticide", "Subject", "Impact"), expand = c(.05, .10)) +
xlab("Variables") +
ylab("Frequency") +
geom_alluvium(aes(fill = subject)) +
geom_stratum(width = 1/2, fill = "white", color = "black") +
geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 5, fontface = "bold") +
theme_minimal() +
theme(axis.text=element_text(size=15),
axis.title=element_text(size=20),
legend.position="none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank()
) +
scale_fill_brewer(palette = "Dark2", name = "Subject Category")
figs31
# ggsave(here("figures", "figs31.pdf"), width =18, height = 12, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs31.jpg"), width = 18, height = 12, units = "cm", scale = 2, dpi = 800)
Figure 4
A bubble plot showing the counts of each pesticide per impact included in current meta-analysis on the impacts of orhanochlorine pesticides
# Join the organochlorine pesticide details with the impact details
ocp_im <- left_join(ocp, im ,sub, by = "study_id")
# Separate rows in "ocp_im"
ocp_im1 <- separate_rows(ocp_im, ocp , sep = ", ", convert = TRUE)
# Group by "ocp" and "impact" and summarize count
ocp_im_summary <- ocp_im1 %>%
mutate(ocp = str_trim(ocp),
impact_broad = str_trim(impact_broad)) %>%
replace_ocp2() %>%
group_by(ocp, impact_broad) %>%
summarise(count = n(), .groups = "drop")
# Filter for top 5 pesticides
top_pesticides <- ocp_im_summary %>%
filter(ocp != "not reported") %>%
group_by(ocp) %>%
summarise(total_count = sum(count)) %>%
top_n(8, total_count) %>%
pull(ocp)
ocp_im_summary_filtered <- ocp_im_summary %>%
filter(ocp %in% top_pesticides)
# Create a circle plot with impact on the x-axis and ocp on the y-axis
fig4 <- ocp_im_summary_filtered %>%
ggplot(aes(x = fct_rev(fct_reorder(impact_broad, count, .fun = 'sum')),
y = fct_reorder(ocp, count, .fun = 'sum'),
size = count,
fill = count)) +
geom_point(shape = 21, color = "white") +
geom_text(aes(label = count), size = 8, fontface = "bold") +
scale_fill_gradient(low = "#98FB98", high = "#006400") +
labs(x = "Impact",
y = "Organochlorine Pesticide",
fill = "Count") +
theme_minimal() +
theme(
axis.ticks.y = element_blank(),
axis.text.x = element_text(size = 20),
axis.text.y = element_text(size = 20, hjust = 1),
axis.title.x = element_text(size = 25),
axis.title.y = element_text(size = 25),
legend.position = "none") +
scale_size_continuous(range = c(15, 45))
fig4
# ggsave(here("figures", "fig4.pdf"), width = 25, height = 15, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "fig4.jpg"), width = 25, height = 15, units = "cm", scale = 2, dpi = 800)
Objective 3
To investigate global research output and collaboration networks
Figure s32, s33, s34, s35 & s36
figs32 <- biblioAnalysis(bib_sco)
plot(figs32)
summary(figs32)
##
##
## MAIN INFORMATION ABOUT DATA
##
## Timespan 1993 : 2022
## Sources (Journals, Books, etc) 45
## Documents 100
## Annual Growth Rate % 9.25
## Document Average Age 7.78
## Average citations per doc 58.6
## Average citations per year per doc 6.523
## References 7548
##
## DOCUMENT TYPES
## article 50
## conference paper 2
## review 48
##
## DOCUMENT CONTENTS
## Keywords Plus (ID) 1592
## Author's Keywords (DE) 258
##
## AUTHORS
## Authors 544
## Author Appearances 684
## Authors of single-authored docs 1
##
## AUTHORS COLLABORATION
## Single-authored docs 1
## Documents per Author 0.184
## Co-Authors per Doc 6.84
## International co-authorships % 41
##
##
## Annual Scientific Production
##
## Year Articles
## 1993 1
## 1995 1
## 1999 1
## 2000 1
## 2004 2
## 2006 3
## 2007 1
## 2008 2
## 2009 1
## 2010 4
## 2011 3
## 2012 5
## 2013 5
## 2014 9
## 2015 7
## 2016 11
## 2017 4
## 2018 2
## 2019 7
## 2020 8
## 2021 9
## 2022 13
##
## Annual Percentage Growth Rate 9.25
##
##
## Most Productive Authors
##
## Authors Articles Authors Articles Fractionalized
## 1 LISON D 7 LISON D 2.000
## 2 VAN MAELE-FABRY G 7 VAN MAELE-FABRY G 2.000
## 3 BONDE JP 5 DAVIS WJ 1.000
## 4 BALLESTER F 4 GAMET-PAYRASTRE L 0.867
## 5 CHEVRIER C 4 KREWSKI D 0.867
## 6 EGGESBØ M 4 HOET P 0.833
## 7 GOVARTS E 4 FU X 0.750
## 8 SCHOETERS G 4 MUÑOZ CC 0.750
## 9 TRNOVEC T 4 VERMEIREN P 0.750
## 10 ZHANG Y 4 LEVY LS 0.700
##
##
## Top manuscripts per citations
##
## Paper DOI TC TCperYear NTC
## 1 BROWN TP, 2006, ENVIRON HEALTH PERSPECT 10.1289/ehp.8095 329 18.28 1.85
## 2 GOVARTS E, 2012, ENVIRON HEALTH PERSPECT 10.1289/ehp.1103767 228 19.00 1.81
## 3 PEZZOLI G, 2013, NEUROLOGY 10.1212/WNL.0b013e318294b3c8 207 18.82 2.08
## 4 BONDE JP, 2016, HUM REPROD UPDATE 10.1093/HUMUPD/DMW036 188 23.50 2.24
## 5 RIGÉT F, 2010, SCI TOTAL ENVIRON 10.1016/j.scitotenv.2009.07.036 162 11.57 1.50
## 6 VAN DER MARK M, 2012, ENVIRON HEALTH PERSPECT 10.1289/ehp.1103881 161 13.42 1.28
## 7 OJAJÄRVI IA, 2000, OCCUP ENVIRON MED 10.1136/oem.57.5.316 153 6.38 1.00
## 8 SCHINASI L, 2014, INT J ENVIRON RES PUBLIC HEALTH 10.3390/ijerph110404449 148 14.80 2.90
## 9 SONG Y, 2016, J DIABETES 10.1111/1753-0407.12325 140 17.50 1.67
## 10 ADAMI HO, 1995, CANCER CAUSES CONTROL 10.1007/BF00054165 140 4.83 1.00
##
##
## Corresponding Author's Countries
##
## Country Articles Freq SCP MCP MCP_Ratio
## 1 CHINA 17 0.1753 12 5 0.294
## 2 USA 11 0.1134 9 2 0.182
## 3 BELGIUM 7 0.0722 5 2 0.286
## 4 CANADA 6 0.0619 3 3 0.500
## 5 FRANCE 6 0.0619 2 4 0.667
## 6 BRAZIL 5 0.0515 5 0 0.000
## 7 DENMARK 5 0.0515 0 5 1.000
## 8 SPAIN 5 0.0515 0 5 1.000
## 9 NETHERLANDS 4 0.0412 3 1 0.250
## 10 UNITED KINGDOM 4 0.0412 3 1 0.250
##
##
## SCP: Single Country Publications
##
## MCP: Multiple Country Publications
##
##
## Total Citations per Country
##
## Country Total Citations Average Article Citations
## 1 DENMARK 687 137.4
## 2 USA 686 62.4
## 3 CHINA 607 35.7
## 4 UNITED KINGDOM 567 141.8
## 5 BELGIUM 488 69.7
## 6 CANADA 417 69.5
## 7 FRANCE 389 64.8
## 8 NETHERLANDS 259 64.8
## 9 ITALY 207 207.0
## 10 SPAIN 181 36.2
##
##
## Most Relevant Sources
##
## Sources Articles
## 1 ENVIRONMENTAL HEALTH PERSPECTIVES 10
## 2 SCIENCE OF THE TOTAL ENVIRONMENT 9
## 3 ENVIRONMENT INTERNATIONAL 7
## 4 CANCER CAUSES AND CONTROL 5
## 5 ENVIRONMENTAL RESEARCH 5
## 6 ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH 5
## 7 ENVIRONMENTAL SCIENCE AND TECHNOLOGY 4
## 8 SCIENTIFIC REPORTS 4
## 9 CHEMOSPHERE 3
## 10 ENVIRONMENTAL POLLUTION 3
##
##
## Most Relevant Keywords
##
## Author Keywords (DE) Articles Keywords-Plus (ID) Articles
## 1 META-ANALYSIS 43 ENVIRONMENTAL EXPOSURE 92
## 2 PESTICIDES 28 HUMAN 84
## 3 SYSTEMATIC REVIEW 18 PESTICIDE 79
## 4 OCCUPATIONAL EXPOSURE 8 HUMANS 71
## 5 DDT 7 FEMALE 70
## 6 CHILD 6 META ANALYSIS 64
## 7 DDE 5 PESTICIDES 60
## 8 BREAST CANCER 4 OCCUPATIONAL EXPOSURE 57
## 9 INSECTICIDES 4 MALE 56
## 10 ORGANOCHLORINES 4 PRIORITY JOURNAL 53
Figure s37
Thematic map based on keywords extracted from ID field (obtained on the Scopus scientific literature database) of meta-analysis included in the systematic review map.
par(mfrow=c(1,1), mar=c(0,2,0,2))
fig37 <- thematicMap(bib_sco, field = "ID", n = 1500, minfreq = 1, stemming = FALSE, size = 1, n.labels = 1, repel = TRUE)
plot(fig37$map)
Figure 5
Heat map of world showing the country-level counts for first authors’ country of affiliation of meta-analysis investigating the impacts of organochlorine pesticides. Grey indicates no publications affiliated with a given country in our data set.
# Extract country information from the "AU1_CO" and "AU_CO" fields of the "bib_sco" dataset
bibmap <- metaTagExtraction(bib_sco, Field = "AU1_CO", sep = ";")
bibmap <- metaTagExtraction(bibmap, Field = "AU_CO", sep = ";")
# Create a data frame with counts of articles from each country
firstcountrycounts <- bibmap %>%
group_by(AU1_CO) %>%
count() %>%
filter(!is.na(AU1_CO))
# Load world map data and remove countries with longitude >180 to make an equal projection-like map
world_map <- map_data("world") %>%
filter(! long > 180)
# Format country names to match regions on the world map
firstcountrycounts$region <- str_to_title(firstcountrycounts$AU1_CO)
firstcountrycounts$region[firstcountrycounts$region == "Usa"] <- "USA"
firstcountrycounts$region[firstcountrycounts$region == "Korea"] <- "South Korea"
firstcountrycounts$region[firstcountrycounts$region == "United Kingdom"] <- "UK"
# Join count data with map data and set missing counts to zero
emptymap <- tibble(region = unique(world_map$region), n = rep(0,length(unique(world_map$region))))
fullmap <- left_join(emptymap, firstcountrycounts, by = "region")
fullmap$n <- fullmap$n.x + fullmap$n.y
fullmap$n[is.na(fullmap$n)] <- 0
fig5 <- fullmap %>%
ggplot(aes(fill = n, map_id = region)) +
geom_map(map = world_map, color = "gray50") +
expand_limits(x = world_map$long, y = world_map$lat) +
coord_map("mercator") +
theme_minimal() +
theme(
axis.text = element_blank(),
axis.title = element_blank(),
legend.position = "bottom",
legend.box = "horizontal",
legend.box.just = "center",
legend.margin = margin(t = 10, unit = "pt"),
legend.text = element_text(size = 12),
legend.title = element_text(size = 14, face = "bold"),
legend.key.width = unit(30, "mm")
) +
scale_fill_gradient(
low = "#98FB98", high = "#006400",
name = "Score", na.value = "gray70",
limits = c(1, 20)) +
guides(
fill = guide_colourbar(
barwidth = unit(180, units = "mm"),
barheight = unit(3, units = "mm")
)
)
fig5
# ggsave(here("figures", "fig5.pdf"), width = 21, height = 12, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "fig5.jpg"), width = 21, height = 12, units = "cm", scale = 2, dpi = 800)
Figure s43
Heat map of Europe showing the country-level counts for first authors’ country of affiliation of meta-analysis investigating the impacts of organochlorine pesticides. Grey indicates no publications affiliated with a given country in our data set.
figs43 <- fullmap %>%
ggplot(aes(fill = n, map_id = region)) +
geom_map(map = world_map, color = "gray50") +
coord_map("mercator", ylim = c(30, 65), xlim = c(-35, 55)) +
theme(
axis.text = element_blank(),
axis.title = element_blank(),
legend.box = "horizontal",
legend.box.just = "center",
legend.margin = margin(t = 10, unit = "pt"),
legend.text = element_text(size = 12),
legend.title = element_text(size = 14, face = "bold"),
legend.key.width = unit(30, "mm"),
legend.position = "bottom"
) +
scale_fill_gradient(low = "#98FB98", high = "#006400",
name = "Score", na.value = "gray70",
limits = c(1, 10)) +
guides(fill = guide_colourbar(barwidth = unit(180, units = "mm"), barheight = unit(3, units = "mm")))
figs43
# ggsave(here("figures", "figs43.pdf"), width = 21, height = 12, units = "cm", scale = 2, dpi = 800)
# ggsave(here("figures", "figs43.jpg"), width = 21, height = 12, units = "cm", scale = 2, dpi = 800)
Figure s44
Chord diagram of collaborations across countries. Countries represent the location of the primary authors’ affiliated institution.
# Extract countries from the affiliations
bib_sco2 <- metaTagExtraction(bib_sco, Field = "AU_CO", sep = ";")
# Create a network matrix of collaborations between countries
NetMatrix_country <- biblioNetwork(bib_sco2, analysis = "collaboration", network = "countries", sep = ";")
# Convert the network matrix to a standard matrix
NetMatrix_country <- as.matrix(NetMatrix_country)
# Remove the lower triangle (as this is duplication of info)
NetMatrix_country[lower.tri(NetMatrix_country)] <- 0
# Change column and row names to title case
colnames(NetMatrix_country) <- str_to_title(colnames(NetMatrix_country))
rownames(NetMatrix_country) <- str_to_title(rownames(NetMatrix_country))
# Change "Usa" to "USA"
colnames(NetMatrix_country)[colnames(NetMatrix_country) == "Usa"] <- "USA"
rownames(NetMatrix_country)[rownames(NetMatrix_country) == "Usa"] <- "USA"
# Change "United Kingdom" to "UK" for easier plotting
colnames(NetMatrix_country)[colnames(NetMatrix_country) == "United Kingdom"] <- "UK"
rownames(NetMatrix_country)[rownames(NetMatrix_country) == "United Kingdom"] <- "UK"
# Setting up custom parameters
circos.par(cell.padding = c(0, 0, 0, 0), track.margin = c(0, 0))
# Create a chord diagram of the network matrix
figs44 <- chordDiagram(NetMatrix_country, annotationTrack = "grid", preAllocateTracks = 1)
# Add a track to label each sector with its name
circos.trackPlotRegion(track.index = 1, bg.border = NA, panel.fun = function(x, y) {
xlim = get.cell.meta.data("xlim")
ylim = get.cell.meta.data("ylim")
sector.name = get.cell.meta.data("sector.index")
circos.text(mean(xlim), ylim[1] + 0.2 , sector.name, facing = "clockwise", niceFacing = TRUE, adj = c(0, 0))
circos.axis(h = "top", labels.cex = 0.5, major.tick.length = 0.2, sector.index = sector.name, track.index = 2)
})
Figure s45
Chord diagram illustration of collaborations across countries. Countries represent the location of the primary authors’ affiliated institution. Collaborations within countries are not shown.
diag(NetMatrix_country) <- 0
# Create a chord diagram of the network matrix
figs45 <- chordDiagram(NetMatrix_country, annotationTrack = "grid", preAllocateTracks = 1)
# Add a track to label each sector with its name
circos.trackPlotRegion(track.index = 1, panel.fun = function(x, y) {
xlim = get.cell.meta.data("xlim")
ylim = get.cell.meta.data("ylim")
sector.name = get.cell.meta.data("sector.index")
circos.text(mean(xlim), ylim[1] + 0.2, sector.name, facing = "clockwise", niceFacing = TRUE, adj = c(0, 0.5))
circos.axis(h = "top", labels.cex = 0.5, major.tick.length = 0.2, sector.index = sector.name, track.index = 2)
}, bg.border = NA)
Figure s46
Chord diagram illustration of collaborations across continents. Continents represent the location of the primary authors’ affiliated institution. Collaborations within countries are not shown.
# Country rename to continents
NetMatrix_continent <- NetMatrix_country
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "USA"] <- "North\nAmerica"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "USA"] <- "North\nAmerica"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Spain"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Spain"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "China"] <- "Asia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "China"] <- "Asia"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "France"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "France"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Canada"] <- "North\nAmerica"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Canada"] <- "North\nAmerica"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Denmark"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Denmark"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Netherlands"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Netherlands"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Belgium"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Belgium"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "UK"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "UK"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Australia"] <- "Australia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Australia"] <- "Australia"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Brazil"] <- "South America"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Brazil"] <- "South America"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Germany"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Germany"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Finland"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Finland"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Greece"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Greece"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Korea"] <- "Asia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Korea"] <- "Asia"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Norway"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Norway"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Sweden"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Sweden"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Egypt"] <- "Africa"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Egypt"] <- "Africa"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Italy"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Italy"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Japan"] <- "Asia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Japan"] <- "Asia"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Portugal"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Portugal"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Switzerland"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Switzerland"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Turkey"] <- "Asia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Turkey"] <- "Asia"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "India"] <- "Asia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "India"] <- "Asia"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Iran"] <- "Asia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Iran"] <- "Asia"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Costa Rica"] <- "North\nAmerica"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Costa Rica"] <- "North\nAmerica"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Czech Republic"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Czech Republic"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Hong Kong"] <- "Asia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Hong Kong"] <- "Asia"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Iceland"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Iceland"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Ireland"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Ireland"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Mexico"] <- "North\nAmerica"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Mexico"] <- "North\nAmerica"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Romania"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Romania"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Slovakia"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Slovakia"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Ukraine"] <- "Europe"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Ukraine"] <- "Europe"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Australia"] <- "Oceania"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Australia"] <- "Oceania"
colnames(NetMatrix_continent)[colnames(NetMatrix_continent) == "Kazakhstan"] <- "Asia"
rownames(NetMatrix_continent)[rownames(NetMatrix_continent) == "Kazakhstan"] <- "Asia"
# collapsing
merge_matrix <- t(rowsum(t(NetMatrix_continent), group = colnames(NetMatrix_continent), na.rm = T))
merge_matrix2 <- rowsum(merge_matrix, group = rownames(merge_matrix))
# remove diagonal elements
diag(merge_matrix2) <- 0
# chord plot
chordDiagramFromMatrix(merge_matrix2)
# Create a chord diagram of the network matrix
figs46<- chordDiagram(merge_matrix2, annotationTrack = "grid", preAllocateTracks = 1)
# Add a track to label each sector with its name
circos.trackPlotRegion(track.index = 1, panel.fun = function(x, y) {
xlim = get.cell.meta.data("xlim")
ylim = get.cell.meta.data("ylim")
sector.name = get.cell.meta.data("sector.index")
circos.text(mean(xlim), ylim[1] + 0.2, sector.name, facing = "clockwise", niceFacing = TRUE, adj = c(0, 0.5))
circos.axis(h = "top", labels.cex = 0.5, major.tick.length = 0.2, sector.index = sector.name, track.index = 2)
}, bg.border = NA)
Figure s47
Chord diagram illustration of collaborations across disciplines. Disciplines have been allocated based on the Journal Citation Categories on Web of Science. Collaborations within disciplines are not shown.
fields <- sd %>%
mutate(
title = str_to_lower(paper_title),
ntitle = paste(str_to_lower(str_split_fixed(study_id, "\\_", n = 2)[,1]), paper_title, sep = " "),
ntitle = str_squish(ntitle),
Journal_Category_Allocated_Broad = str_replace_all(Journal_Category_Allocated_Broad, "Environmental Science", "Environmental\nScience ")
) %>%
select(ntitle, Journal_Category_Allocated_Broad)
bib_names <- bib_sco %>% rownames_to_column(., var = "mat_names") %>%
mutate(TI2 = tolower(unlist(lapply(data.frame(t(str_split_fixed(TI, " ", n = 15)[,1:14])),
function(x) str_c(x, collapse = " ")))),
name2 = stri_trans_general(tolower(str_split_fixed(SR, " ", n = 2)[,1]), "latin-ascii"),
TI2 = paste(name2, TI2, sep= " "),
TI2 = trimws(TI2)) %>%
select(TI2, mat_names)
pos <- lapply(bib_names$TI2, function(x) stringdist(fields$ntitle, x))
pos2<- map_dbl(pos, which.min)
# now we can merge two datasets
bib_names$Journal_Category_Allocated_Broad <- fields$Journal_Category_Allocated_Broad[pos2]
# Creating matrix for bibliometric coupling
NetMatrix <- biblioNetwork(bib_sco, analysis = "coupling", network = "references", sep = ";")
# forcing into a nromal matrix
net_matrix <- as.matrix(NetMatrix)
diag(net_matrix) <- 0 #get rid of counts for the same papers
# replacing names with Journal_Category_Allocated_Broad
rownames(net_matrix) <- bib_names$Journal_Category_Allocated_Broad
colnames(net_matrix) <- bib_names$Journal_Category_Allocated_Broad
# reducing matrix according to discipline_code
rect_matrix<- t(rowsum(t(net_matrix), group = colnames(net_matrix), na.rm = T))
small_matrix <- rowsum(rect_matrix, group = rownames(rect_matrix))
# getting rid of lower triangle (as this is duplication of info)
small_matrix[lower.tri(small_matrix)] <- 0
par(mar = c(0, 0, 0, 0), mfrow = c(1, 1))
# Create a chord diagram of the network matrix
figs47 <- chordDiagram(small_matrix, annotationTrack = "grid", preAllocateTracks = 1)
# Add a track to label each sector with its name
circos.trackPlotRegion(track.index = 1, panel.fun = function(x, y) {
xlim = get.cell.meta.data("xlim")
ylim = get.cell.meta.data("ylim")
sector.name = get.cell.meta.data("sector.index")
circos.text(mean(xlim), ylim[1] + 0.2, sector.name, facing = "clockwise", niceFacing = TRUE, adj = c(0, 0.5))
circos.axis(h = "top", labels.cex = 0.5, major.tick.length = 0.2, sector.index = sector.name, track.index = 2)
}, bg.border = NA)